Re: [RFC][PATCH] Make cryptoapi non-optional?

[Matt Mackall]

On Sat, Aug 09, 2003 at 07:07:27PM -0700, Robert Love wrote:
> On Sat, 2003-08-09 at 07:33, Matt Mackall wrote:
> 
> > - the random number generator is non-optional because it's used
> >   various things from filesystems to networking
> 
> What if you kept crypto API optional, made random.c a config option, and
> make that depend on the crypto API. Then -- and this is the non-obvious
> part -- implement a super lame replacement for get_random_bytes() [what
> I assume the various parts of the kernel are using] for when
> !CONFIG_RANDOM is not set?
> 
> You can do a simple PRNG in <10 lines of C. Have the kernel seed it on
> boot with xtime or something else lame.

Eek, let's not go there. Most of the users of get_random_bytes()
depend on it being strong enough for cryptographic purposes. And we've
long guaranteed the existence of a cryptographically strong
/dev/urandom. If userspace could no longer rely on it being there,
people would be forced to go back to poorly reinventing it in
userspace - not good. Having random.c be nonoptional is a good note to
embedded folks that they should think long and hard before ripping it
out.

-- 
Matt Mackall : http://www.selenic.com : of or relating to the moon
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/